A&A 633, A93 (2020) Astronomy https://doi.org/10.1051/0004-6361/201936804 & © ESO 2020 Astrophysics Formation of moon systems around giant planets Capture and ablation of planetesimals as foundation for a pebble accretion scenario T. Ronnet and A. Johansen Department of Astronomy and Theoretical Physics, Lund Observatory, Lund University, Box 43, 22100 Lund, Sweden e-mail: [email protected] Received 28 September 2019 / Accepted 10 December 2019 ABSTRACT The four major satellites of Jupiter, known as the Galilean moons, and Saturn’s most massive satellite, Titan, are believed to have formed in a predominantly gaseous circum-planetary disk during the last stages of formation of their parent planet. Pebbles from the protoplanetary disk are blocked from flowing into the circumplanetary disk by the positive pressure gradient at the outer edge of the planetary gap, so the gas drag assisted capture of planetesimals should be the main contributor to the delivery of solids onto circum- planetary disks. However, a consistent framework for the subsequent accretion of the moons remains to be built. Here, we use numerical integrations to show that most planetesimals that are captured within a circum-planetary disk are strongly ablated due to the frictional heating they experience, thus supplying the disk with small dust grains, whereas only a small fraction “survives” their capture. We then constructed a simple model of a circum-planetary disk supplied by ablation, where the flux of solids through the disk is at equilibrium with the ablation supply rate, and we investigate the formation of moons in such disks. We show that the growth of satellites is mainly driven by accretion of the pebbles that coagulate from the ablated material. The pebble-accreting protosatellites rapidly migrate inward and pile up in resonant chains at the inner edge of the circum-planetary disk. We propose that dynamical instabilities in these resonant chains are at the origin of the different architectures of Jupiter’s and Saturn’s moon systems. The assembly of moon systems through pebble accretion can therefore be seen as a down-scaled manifestation of the same process that forms systems of super-Earths and terrestrial-mass planets around solar-type stars and M-dwarfs. Key words. planets and satellites: individual: Galilean moons – planets and satellites: individual: Titan – planets and satellites: gaseous planets – planets and satellites: formation 1. Introduction Finally, information regarding the internal structure of the moons has been inferred from the gravity measurements performed The formation of massive moons around gas giant planets is by the Galileo spacecraft at the Jovian system and the Cassini envisioned to take place in a gaseous disk surrounding the planet spacecraft in the case of Saturn’s moons. While the three inner in the last stages of its accretion (see, e.g., Peale & Canup 2015, Galilean moons (involved in the Laplace resonance system) are for a review). Their formation would thus be analogous to that likely to be fully differentiated with the presence of an iron core, of planets around stars and face the same theoretical challenges. a silicate mantle and, in the case of Europa and Ganymede, an icy These difficulties include, for example, the formation of satel- outer mantle, Callisto and Titan appear to be only partially differ- litesimals to seed the accretion of more massive moons and entiated (Anderson et al. 2001; Iess et al. 2010). The formation of the survival of these latter against their rapid inward migra- undifferentiated satellites implies limited heating during accre- tion. In addition, some issues, such as the delivery of solids to tion to prevent large scale ice melting and hence requires long the circum-planetary disk (hereafter CPD), are specific to the formation timescales (&1 Myr; Barr & Canup 2008). Even so, a formation of satellites around the giant planets. The observed later differentiation of the satellites could be difficult to avoid if properties of the Galilean moons orbiting Jupiter, as well as compositional gradients prevent an efficient transport of radio- that of Saturn’s moons, provide additional constraints on the genic heating through convection (O’Rourke & Stevenson 2014). conditions under which they have formed. Also, nonhydrostatic effects could have an important impact on The satellite systems of Jupiter and Saturn both represent a the derivation of the moment of inertia of slowly rotating satel- 4 similar fraction of the mass of their parent planets ( 10− Mp) lites, such as Callisto and Titan (Gao & Stevenson 2013), so and are quite compact (e.g., Canup & Ward 2006). In∼ the case that the interpretation of their internal structure remains uncer- of the Galilean system, the three inner moons – Io, Europa and tain and these objects might in fact be differentiated. If this Ganymede – form a resonant chain (known as a Laplace reso- is the case, these satellites could have formed over much shorter nant system). Additionally, the decreasing bulk densities of the timescales. Galilean satellites with respect to their distance from Jupiter The consideration of the properties of the satellite systems, is suggestive of a compositional gradient among the moons, together with the available knowledge on processes relevant to that is, an increasingly volatile rich composition with respect planet formation, guided the development and refinement of to the distance from Jupiter (see, e.g., Hussmann et al. 2015). satellite formation scenarios (e.g., Lunine & Stevenson 1982; Article published by EDP Sciences A93, page 1 of 20 A&A 633, A93 (2020) Coradini et al. 1995; Canup & Ward 2002, 2006; Mosqueira stop the migration of the satellites (Sasaki et al. 2010; Ogihara & & Estrada 2003a,b; Mousis et al. 2002; Sasaki et al. 2010). Ida 2012). It thus seems that a gas-starved environment is not Although the gas-starved model developed by Canup & Ward essential to allow for the formation of icy satellites and their (2002, 2006) has been well recognized as a plausible satellite for- survival. mation scenario, in recent years, new paradigms have emerged On the other hand, one of the important issues of the gas- regarding several key processes for the formation of planets, starved model is the formation of satellitesimals and satellites which challenge our current understanding of the formation of seeds. Whereas it has been customarily assumed that large the giant planets’ moons. objects would form out of the small dust grains in the CPD (e.g., These challenges are briefly discussed in Sect.2, where we Canup & Ward 2002, 2006; Sasaki et al. 2010; Ogihara & Ida argue that the capture and ablation of planetesimals should be an 2012), it is now understood that the formation of planetesimals important source of solids in giant planet’s CPD, as previously or satellitesimals likely requires some instability (e.g., streaming suggested (e.g., Estrada et al. 2009), but unlike the assumption instabilities) or adequate environment (e.g., pressure bumps) to of the gas-starved model. Such a mechanism is investigated in allow for the efficient concentration of dust that can then grav- Sect.4, which shows that most of the planetesimals captured in itationally collapse into 10–100 km sized objects (see Johansen the CPD are efficiently ablated. Ultimately, the material ablated et al. 2014, for a review). The rapid inward drift timescale of off of the surface of the planetesimals provides a source of dust dust grains in a CPD precludes the formation of satellitesimals in the CPD whose subsequent evolution is investigated in Sect.5. through any known mechanism (Shibaike et al. 2017). Even if These results provide the ground for the development of a revised satellitesimals were somehow able to form (see, for example, 1 formation model for the giant planets’ satellites (Sect.6). Specif- Dr˛a˙zkowska & Szulágyi 2018) , it is not expected that all the ically, we propose that the seeds of the satellites initially form dust grains would be converted into large objects, contrary to the from the fraction of captured planetesimals that survived abla- assumption in previous investigations. Rather, the satellitesimals tion in the CPD and subsequently grow through pebble accretion should further grow by accreting the remaining (and possibly from the flux of dust supplied by the ablation of planetesimals. re-supplied) dust grains in a process known as pebble accre- More aspects and implications are discussed in Sect.7, before tion (see the recent reviews by Johansen & Lambrechts 2017; we summarize our results in Sect.8. Ormel 2017). The accretion of moons in CPDs thus needs to be reassessed considering this more likely growth channel. Another important issue is that it is unlikely that the material 2. Motivation accreted by a giant planet in the late stages of its formation has a solar dust-to-gas mass ratio, as advocated in the gas-starved The so-called gas-starved model developed by Canup & Ward models (see Ronnet et al. 2018, for a discussion). Above a mass (2002, 2006) has been the leading scenario for the formation of of typically a few 10 M , a protoplanet starts to significantly satellites around gas giant planets because of its apparent ability perturb the gas disk∼ in its⊕ vicinity, opening a shallow gap whose to resolve long-standing issues which are primarily (i) the impos- outer edge acts as a barrier for drifting dust grains (Morbidelli & sibility of satisfying the minimum mass requirement to grow the Nesvorny 2012; Lambrechts et al. 2014). At a mass compara- moons and the low temperature needed for ice stability simul- ble to that of Jupiter, dust grains larger than a few 10–100 µm taneously in a “self-consistent” viscous disk model (Makalkin should be efficiently filtered (Paardekooper 2007; Zhu et al. et al. 1999; Canup & Ward 2002), (ii) the difficulty of forming 2012; Bitsch et al.